Well, no, actually, but the end result might be the same. This fascinating article in a recent issue of New Scientist magazine looks at the use of crime prediction software in the US, UK and Europe to try to beat criminals at their own game.
Several large software companies have developed programmes which measure statistics and probabilities once stuffed with the relevant data. Some focus on one particular type of crime (burglary, say) while others are more geographically based, but all use a database to try to predict where crimes are most likely to happen.
Surprisingly, it seems to work, although I’d be interested to know whether sheer luck has been factored in or out. The case quoted in the article, for instance, where a police force in Kent was directed to an area they wouldn’t normally have concentrated on and found a violent crime in progress could easily have been down to coincidence.
The article lists various benefits, including the obvious better targeting of limited resources. However, they don’t mention the tendency to rely too much on the software rather than good ‘old fashioned’ information-gathering, which might lead to a “too many eggs in one basket” approach where all the resources are focused on a few areas while the crime pattern is actually more widely spread. Also, could there come a time when the criminals themselves exploit this approach – perhaps by committing lots of minor crimes in one area in order to divert police resources away from their real target somewhere else?
Food for thought; perhaps the police need a crystal ball after all to answer all the questions raised.